AI regulations: USA, China, Europe

In this article, I would like to briefly explain the differences between the approach to AI development in the US, China and Europe, as it is interesting to see how these global players differ from and influence each other in this field. It could show attitude to new technologies in those countries in generall.

USA: The approach of AI regulations in the USA is largely influenced by the National Artificial Intelligence Initiative (NAII) in 2021. The main goals of these regulations are centered around development, education, and building trust in AI systems. One important aspect is the right of explanation, which ensures transparency in AI decision-making processes.

Risk management is another key aspect of the regulations, which is based on three criteria: assessment, independence, and review. This means that AI systems need to undergo thorough evaluations to identify and mitigate potential risks. Additionally, there is a focus on ensuring independence in the evaluation process and establishing mechanisms for regular reviews.

In terms of development, the regulations prioritize granting full autonomy to AI players. This suggests that there is an emphasis on promoting innovation and allowing AI developers to have the freedom to explore and create within the established regulatory framework.

China: The approach of AI regulations in China is driven by the goal of becoming a leader in the field of artificial intelligence. The government supports a collaborative approach, encouraging the participation of “AI National Champions” in advancing AI technology. The regulations primarily focus on three main goals or topics: development, national security, and gaining an international advantage.

In terms of development, there are relatively few restrictions imposed on AI developers, allowing them a considerable amount of freedom to innovate and create. The emphasis is on fostering a supportive environment for AI research and development, enabling rapid technological advancements.

National security is also a key concern, with regulations aiming to ensure that AI technologies do not pose risks or threats to the country’s security. 

Lastly, gaining an international advantage is a priority, suggesting a focus on fostering competitiveness in the global AI market. The regulations likely aim to promote the growth of Chinese AI companies and encourage their expansion and influence beyond national borders.

Europe: The approach of AI regulations in Europe is strongly influenced by the GDPR (General Data Protection Regulation), which plays a significant role in establishing rules. A risk-based approach is followed, categorizing regulations into four levels based on the potential risks associated with AI applications. The priority is to ensure user safety, and AI systems are not allowed to make significant decisions autonomously. The right of explanation is considered highly important, emphasizing the need for transparency in AI decision-making processes.

Regulations address the issue of deepfakes and similar techniques by requiring a disclosure comment to be attached to such content. Social scoring using AI is prohibited, aiming to prevent discriminatory or unfair practices. Certain areas are subject to a ban on AI applications, likely for reasons of ethical concerns or potential risks.

However, this approach has faced criticism for being perceived as “not flexible.” This criticism likely suggests that the regulations may be considered rigid or lacking adaptability to address emerging challenges or technological advancements in the AI field.

These diverse approaches are not isolated but mutually influence each other. The competitive nature of the global AI market and the desire to lead in technological advancements create a dynamic where regions learn from one another’s experiences and adapt their regulations accordingly. Cultural, political, and economic factors play a significant role in shaping these approaches, reflecting the unique values and priorities of each region.

Experts advise that Europe should adopt a more flexible approach to AI regulations while the USA should enhance privacy protections by adopting stricter measures. In order to achieve these goals, several recommendations have been put forth:

These include encouraging the sharing of documentation to enhance transparency, promoting collaborative research efforts, establishing standards for AI development, creating AI sandboxes for safe experimentation, supporting large-scale public AI research projects and open-source tools, facilitating regulator-to-regulator exchanges, and developing an AI assurance ecosystem for accountability and trust. By implementing these recommendations, both regions can foster innovation, ensure responsible AI development, and promote international cooperation and harmonization in AI regulations.

As you can see, there is still a long way to go before the global players arrive at a universal approach, but the first steps have already been taken.

References:

The EU and U.S. diverge on AI regulation: A transatlantic comparison and steps to alignment | Brookings. (2023, 24. Juni). Brookings. https://www.brookings.edu/research/the-eu-and-us-diverge-on-ai-regulation-a-transatlantic-comparison-and-steps-to-alignment/

Grunitz, M. (2021, 17. Dezember). AI regulation: EU, USA & China – WeAreBrain Blog. WeAreBrain Blog. https://wearebrain.com/blog/ai-data-science/ai-regulation-eu-usa-china/

AI Regulation News | European Union | USA | China. (o. D.). https://www.airegulation.news/

Welcome back!

Welcome back to my research blog!

Last semester I studied high technology, attitudes toward it, and good and bad examples of its use.

In the second semester, I would like to take a closer look at the causes and consequences of innovations, both successful and failed. I would also be interested in taking different states around the world and comparing social attitudes, as well as government regulation of innovation and high technology.

Technology has always accompanied human civilization, as well as aspirations for the future. Has the idea of what will happen to humanity in 100 years changed over the generations? I would like to examine this question and compare expectations of the future in 1900 and today.

And at the very end, I would like to suggest an application that I think can serve as a good example for the meaningful use of high technology. However, there are two sides to any coin.

Common Visual Impairments and possible Solutions for UI

Visual Impairments

  • Nearisghtedness (myopia):
  • Blurred distance vision
  • Good near vision
  • Elongation of the eyeball

    Farsightedness (Hyperopia):
  • Difficulty with near vision
  • Better television viewing
  • Short eyeball

    Astigmatism:
  • Distorted or blurred vision
  • Crooked cornea
  • Blurred vision at all distances

    Presbyopia:
  • Difficulty with near vision
  • Reduced near accommodation
  • Age-related loss of elasticity of the crystalline lens

    Glaucoma (glaucoma):
  • Damage to the optic nerve
  • Increased intraocular pressure
  • Gradual loss of vision

    Macular degeneration:
  • Loss of central vision
  • Impaired reading and recognition of details
  • Degenerative changes in the macula

    Diabetic retinopathy:
  • Damage to the blood vessels of the retina
  • Blurred vision
  • Progressive vision loss

    Color blindness (including red-green blindness):
  • Impaired color vision
  • Difficulty distinguishing specific colors
  • Genetic predisposition

    Nystagmus:
  • Uncontrollable eye movements
  • Impaired vision tracking
  • Blurred vision

    Digital Vision Stress Syndrome (CVS):
  • Eye strain from prolonged viewing of screens
  • Dry eyes
  • Headaches and eye fatigue

    Tunnel vision (restriction of peripheral vision):
  • Narrowed field of vision
  • Lack of peripheral situational awareness
  • Focus on the central area

possible Solutions

  • Nearsightedness (myopia):
  • Adjustable font size
  • Magnifiable UI elements
  • Clear contrast between text and background

    Farsightedness (Hyperopia):
  • Clearly legible font
  • Customizable zoom options for UI elements
  • Good distinguishability of near and far information

    Astigmatism:
  • Sharp and clear icons and text
  • Balanced color contrasts
  • Consistent alignment of UI elements

    Presbyopia:
  • Larger fonts
  • Good readability in different lighting conditions
  • Option to adjust focus settings

    Glaucoma (glaucoma):
  • Highlighting of important information
  • Clear and conspicuous warning icons
  • Adjustable brightness settings

    Macular Degeneration:
  • Option to magnify text and icons
  • Alternative displays for detailed information
    Easy navigation and orientation aids

    Diabetic retinopathy:
  • High contrast color palettes
  • Option to customize UI colors
  • Clear icons for important information

    Color blindness (including red-green blindness):
  • Color modes for color blindness
  • Alternatives to color coding of information
  • Use of shapes and patterns to distinguish

    Nystagmus:
  • Stable and clear UI elements
  • Reduced motion and animation
  • highlighting important information

    Digital Vision Stress Syndrome (CVS):
  • Option to adjust brightness and contrasts
  • Option to reduce blue light emissions
  • Regular reminders to take breaks and do eye exercises

    Tunnel vision (limitation of peripheral vision):
  • Focus on key information and warnings
  • Easy accessibility of maps or radar displays
  • Clear and distinct highlighting of targets or enemies

https://gameaccessibilityguidelines.com/

https://www.playstation.com/en-us/games/the-last-of-us-part-ii/accessibility/

https://www.cdc.gov/visionhealth/basics/ced/index.html#:~:text=The%20leading%20causes%20of%20blindness,disorders%20include%20amblyopia%20and%20strabismus.

My Prototype

As mentioned in the first blog post, my prototype consists of a customizable HUD for a racing game. The goal is to make the gaming experience as enjoyable as possible while also accommodating people with limitations.

The prototype was developed for the fictional racing game “Warp Speed”. I focused mainly on the display of individual HUD elements and the navigation of customization options in the game settings. However, this is only part of the basics for user-friendly and accessible games, as explained in the other blog posts.

The prototype was created using Figma. The fictional game is designed for PC, Xbox, and PlayStation, and is controlled with a mouse or controller.

In the “Options” and “HUD” area, the player can go through the individual elements one by one and select the most suitable UI element. On the right side of the screen is the navigation and an overview of the selectable UI elements. In the center of the screen is a thumbnail of a representative game scene where the selected UI elements can be placed. The right side of the screen displays the available elements from which the player can choose. Finally, after the player has selected all the elements, there is an option to drag and drop them and adjust their size.

In this example, all elements are center-aligned, which people with a limited field of vision might prefer:

Idea for underwater composition

The idea of the composition aims to raise awareness about issues of noise pollution in our oceans. This composition is intended to attract listeners’ attention and highlight the harmful impact of anthropogenic activities on marine life. It should represent balance between the beauty of marine ecosystems and the destructive presence of noise produced by human.

The opening of composition introduces undisturbed marine environment. A combination of soft, gentle melodies with soothing rhythms should bring an image of underwater world where sea creatures communicate through harmonious sounds and songs. Listeners immerse themselves in this soundscape and they become aware of how fragile is life in the ocean. [Time: ~ 1.5/2 min]

The next part of the composition introduces dissonant sounds, symbolizing human interference in the marine environment. The noises are coming from ships engines, seismic surveys and offshore construction disturb a peaceful balance of the underwater ecosystem. The composition builds tension through a series of clashing tones and irregular rhythms. [Time: ~1.5 min]

After introducing a decreased level of noise a brief moment of silence will appear as a reminder of the importance of silence to marine life. That’s a moment where silence reinforces the appreciation of the natural acoustic harmony, which helps marine mammals to survive. [Time: ~0.5 min]

The next part of composition captures the confrontation of marine animals with constant noise pollution. The composition combines melancholic melodies with dissonant themes that should represent destructive sounds. Here we can hear whale songs, dolphin clicks in interaction with anthropogenic sounds. This is a sound awakening a sense of emergency. [Time: ~3 min]

At the end the composition changes again from dissonance to harmony, symbolizing an effort to take an action in the ocean. The disturbing sounds slowly fade away coming back to the natural underwater acoustics. This suppose to inspire listeners to protect marine habitats. [Time: ~1.5/2 min]

By blending contrasting musical elements this composition address the issue of noise pollution in our oceans. It invites everyone to take action and ensure a harmony between human activities and marine life, preserving the seas for future generations. 

Studien zur Smounds Wahrnehmung

Die Bewertung der empfundenen Angenehmheit und Qualität von Gerüchen kann durch die gleichzeitige Wahrnehmung von auditiven Reizen beeinflusst werden, das ist mittlerweile wissenschaftlich bewiesen (vgl. Velasco, Balboa, Marmolejo-Ramos, Spence, 2014).

Carlos Velasco, Diana Balboa, Fernando Marmolejo-Ramos und Charles Spence untersuchten darauf aufbauend, ob die hedonische Kongruenz zwischen Geruchs- und Klangreizen die Wahrnehmung der Geruchsintensität, -angenehmheit und -qualität bei untrainierten Teilnehmern beeinflussen kann. “Hedonische Kongruenz” bezieht sich hierbei auf die Übereinstimmung oder Harmonie zwischen verschiedenen Sinnesreizen hinsichtlich ihrer emotionalen Bewertung oder angenehmen Wirkung. Sprich: ob die angenehme oder unangenehme Wirkung eines Geruchs mit der Bewertung der angenehmen oder unangenehmen Wirkung eines gleichzeitig präsentierten Klangs übereinstimmt.

Velasco et. al (2014) verwendeten für ihre Untersuchungen sechs Gerüche: einen aus dem “Le Nez du Café” (Brizard and Co, Dorchester, UK) und fünf aus den “Le Nez du Vin”- Kits (Brizard and Co, Dorchester, UK). Diese Kits wurden entwickelt, um Kaffe- und Weinliebhaber:innen häufig vorkommende Aromen von Kaffe und Wein näherzubringen. Die Gerüche wurden auch schon in anderen Forschungsarbeiten herangezogen, die sich mit kreuzmodalen Assoziationen zu Gerüchen beschäftigen (z. B. Crisinel und Spence, 2012; Hanson-Vaux et al., 2013).

Die Proben wurden basierend auf einer vorangehenden Untersuchung ausgewählt. Zur Anwendung kamen: Zitrone, Orange und Heidelbeere (als angenehm bewertet), sowie

Moschus, dunkle Schokolade und Rauch (als unangenehm bewertet).
Kombiniert wurden die Gerüche jeweils mit konsonanten (als angenehm bewertet) und dissonanten Klängen (als unangenehm bewertet) und mit weißem Rauschen (als unangenehm bewertet).
Für die Wissenschafter überraschend, zeigte sich, dass das weißes Rauschen tatsächlich einen stärkeren Effekt auf die Geruchsbewertungen der Versuchspersonen hatte, als die konsonanten oder dissonanten musikalischen Auswahlstücke. Die Teilnehmer bewerteten die verwendeten Gerüche als weniger angenehm, wenn sie dem weißen Rauschen lauschten. Grundsätzlich schränken Carlos Velasco, Diana Balboa, Fernando Marmolejo- Ramos und Charles Spence die Resultate der Untersuchungen in vielerlei Hinsicht ein, letztlich bleibt als einzige signifikante Erkenntnis, dass Gerüche unangenehmer bewertet werden, wenn gleichzeitig weißes Rauschen zu hören ist.

Quellen:

Hanson-Vaux, G., Crisinel, A. S., and Spence, C. (2013). Smelling shapes: crossmodal correspondences between odors and shapes. Chem. Senses 38, 161–166. doi: 10.1093/ chemse/bjs087

Velasco Carlos, Balboa Diana, Marmolejo-Ramos Fernando, Spence Charles: Crossmodal effect of music and odor pleasantness on olfactory quality perception. Frontiers in Psychology 5, 2014. https://www.frontiersin.org/articles/10.3389/fpsyg.2014.01352


Insta-Sara: Showcasing the Prototype Alongside a Storyboard.

After spending the past few weeks focusing on potential content for an Instagram Youth version, I have now compiled them into a storyboard. The story revolves around a teenager named Sara and her first experiences with Instagram, aiming to demonstrate how certain features can be applied in everyday life. At this stage, the emphasis is less on the app’s design and more on the events surrounding it, showcasing how the functions can be utilized.

Have fun watching it 🙂

Whales vocalization

Human fascination with whale songs goes back to the 1970s, when a biologist Roger Payne recorded an album of humpback whale songs. This inspired Greenpeace’s “Save the Whales” movement and led to an international ban on deep-see whaling. Many whales make sounds but when we talk about whale songs we’re usually talking about humpback whales. Humpback songs are long, complex patterns of sounds that whales sometimes repeat for days on end.

Whale hearing
Since all mammals underwater are especially sensitive to the sounds in order to be able to survive, whales developed their hearing. Looking at the anatomy of whales the size of whale’s head is up to one third of the total body length. The whale ear is very tiny. The bone structure of the middle and inner ears is modified to accommodate hearing underwater.

The Humpback song
Humpbacks don’t breathe like humans do. A humpback whale actually has to think about breathing. To go to sleep and still breathe, they’ll only shut off half their brain at a time. Originally it was thought the sound is produced through larynx, but research showed that larynx doesn’t move during vocalisations. There are some nasal structures which move when sound is produced. The research shows that only male humpback sing and usually the same song. They produce moans, grunts, blasts and shrieks. Since the sound waves can travel very far in water without losing energy, some of the low frequencies can travel approx up till 16 094 km in some levels of the ocean. The frequency range is 30Hz till around 8 kHz, which means that humans are only able to hear only a part of the whales’ songs.

FIN WHALE
These whales make some of the lowest-frequency sounds of any animal. Males produce songs made up of loud pulses lasting 1–2 seconds, in patterns that can last anywhere from several minutes to days

Sound levels

The frequency of baleen whale sounds ranges from 10 Hz to 31 kHz. A list of typical levels is shown in the table below.

SourceBroadband source level (dB re 1 Pa at 1m)
Fin whale moans155–186
Blue whale moans155–188
Gray whale moans142–185
Bowhead whale tonals, moans and song128–189

The “notes” are single uninterrupted emissions of sound that last up to a few seconds. These sounds vary in frequency from 20 Hz to upward of 24 kHz (the typical human range of hearing is 20 Hz to 20 kHz). It may be frequency modulated or amplitude modulated. However, the adjustment of bandwidth on a spectrogram representation of the song reveals the essentially pulsed nature of the FM sounds.

References:
1. Ocean Noise – from science to management – Journal of marine science and engineering
2. https://en.wikipedia.org/wiki/List_of_whale_vocalizations
3. https://oceanservice.noaa.gov/facts/whalesounds.html

Basics of Accessible Video Game Design

  • About 20% of the population has a disability.
  • The objective of accessible design is to eliminate barriers and ensure equal opportunities in gameplay.
  • Accessibility is particularly crucial for players with disabilities, visual impairments, hearing loss, or injuries.
  • Different players have different needs and preferences, so there is no one-size-fits-all solution. Therefore, a customizable heads-up display (HUD) is also beneficial.

Visual elements:

  • Individuals have varying visual impairments, screen sizes, and viewing distances. Therefore, it is important to allow players to adjust the text size.
  • Approximately 300 million people have color vision deficiencies, so essential information/feedback should not rely solely on colors.
  • Moreover, contrast should be individually adjustable for players. This is particularly important because elements in dark games may be difficult to recognize in bright lighting conditions.
  • Including haptic feedback, such as vibration or acoustic cues, is also helpful, as it can assist players with visual impairments. By conveying feedback through different senses, accessibility is improved.
  • For subtitles, care should be taken to ensure appropriate speed and length, with no more than 37 characters per line.
  • Subtitles and captions should be large, a simple font should be used and contrast with the background.
  • Additionally, subtitles should be individually adjustable in terms of size and color.
  • For the hearing impaired, it is important to subtitle sounds, such as the footsteps of an approaching opponent.
  • For individuals with epilepsy, there should be an option to disable flashing lights and repetitive patterns.

Auditory elements:

  • Once again, it is important to provide visual or haptic feedback in addition to audio feedback to enhance the gaming experience for the hearing impaired.
  • Separate volume controls for music, speech, and sound effects should be offered, as this allows hearing impaired individuals to adjust the audio elements accordingly.
  • Background noise should be minimized when dialogue takes place in the game.

Controls:

  • When it comes to controls, offering as many options as possible is advantageous. This includes controller, mouse and keyboard, eye-tracking, voice control, etc. Different people with physical impairments can then choose the control option that suits them best.
  • The control of individual game actions should also be individually adjustable.
  • The movement radius and sensitivity of the mouse should be changeable in the settings.
  • For individuals with disabilities, alternative options to quick-time events where certain keys need to be pressed repeatedly should be provided.

https://uxdesign.cc/accessible-video-game-design-7f54c583a470

https://www.playstation.com/en-us/games/the-last-of-us-part-ii/accessibility/

Goals for this Semester

At first, my aim was to make a list of common UI elements that are typically seen in the HUDs of popular gaming genres. The idea was to gather various ways these elements can be displayed. By using practical examples from well-known games, I wanted to show different visual options and figure out which ones work well and which ones don’t. During my research, I realized that the definition of good or bad design is subjective and influenced by personal taste and preferences. For some people, larger elements are important as they enable better readability, while others prefer smaller UI elements because they believe larger ones can be distracting and divert attention from the game itself.

A solution that provides everyone with the opportunity to have a perfect HUD is to make it customizable according to individual preferences People with visual impairments would especially benefit from this.

For my prototype, I will develop a customizable HUD for a racing game. I will present common UI components in various variations. The user will be able to choose their preferred option and adjust the size and color contrast on their own. Then, I will demonstrate different ways the HUD can be displayed, which will look different based on the users’ specific needs. For example, the HUD of a user with red-green color blindness will be different from that of someone with tunnel vision.