How does AI Ethics help businesses go beyond doing what is ethical?

Jan 3, 2024

AI Ethics: Going Beyond What is Right

In December 2023, following two years of investigation by the National Highway Traffic Safety Administration (NHTSA), TESLA recalled over 2 million vehicles on U.S. roads. 

Covering nearly all their vehicles, the largest-ever TESLA recall occurred after NHTSA cited safety concerns over TESLA’s autopilot driver-assistance system. Where it was found to be insufficient in preventing driver abuse and could increase the risk of a crash [1]. Undoubtedly, the costs for such recall – both monetary and reputational – are monumental. And despite TESLA’s disagreement to NHTSA’s analysis, the NHTSA had found (among many) that “drivers are not always paying attention when the autopilot system is on”.

This example, one of many, demonstrates the importance of prioritizing and supporting Artificial Intelligence (AI) Ethics as part of AI adoption or development plans for businesses. Against the background of “unforeseen” AI technology misuse, perhaps in 2024, businesses would start to seriously reconsider the value of having a qualified AI Ethics team and framework to consult prior to their AI deployment.

Commonly placed under “ex-ante” approach to AI Ethics assessment, a qualified AI Ethics team and framework could assist businesses in identifying human misuses before their AI deployment. Especially during the designing and integration stage. In the process, help businesses to reduce possibilities of unnecessary business expenditures such as recalling, reputational damage, compliance risks, and lawsuits.

Additionally, human “misuse” of AI is not exclusive to TESLA. According to OECD AI Incidents Monitor (AIM), between October 2023 to December 2023, there were about 1763 incidents or article coverage of events related to AI [2].

One of the most recent editions to OECD AIM’s is on a ruling by the UK Supreme Court that rejected AI output as “patent worthy”. A ruling adjudicated after a US computer scientist attempted to profit from patenting AI output [3]. Albeit, this ruling is not necessarily considered as a “high-profile” incident, it is nonetheless a ruling that extends the theme of unforeseen misuse of AI by humans.

Unfortunately, this theme has not been appropriately addressed by businesses when we reflect decisions made by tech-giants in disbanding, devaluing, or completely dismissing their AI Ethics team and leadership over the past few years [4]–[6]

Considering these AI related incidents, it is perhaps possible in 2024, AI Ethics in its entirety could be “revalued” in similar vein of importance with all other AI developments and achievements. As AI becomes more ubiquitous in parallel with upcoming AI regulations, and with growing customer and market demand for ethical AI, it is argued here that businesses will have a tough road ahead if they believe AI Ethics as merely a philosophical discipline without any bearing on business profitability.

Ahmad Firdaus – AI Ethicist at Verdas AI

REFERENCES

1] David Shepardson, ‘Tesla recalls nearly all vehicles on US roads over lack of Autopilot safeguards’, Reuters, December 14 20223. [Online]. Available: https://www.reuters.com/business/autos-transportation/tesla-update-software-autopilot-control-issue-2-mln-vehicles-nhtsa-2023-12-13/

[2] ‘OECD AI Incidents Monitor (AIM) October 2023 – December 2023’. OECD, Dec. 28, 2023. [Online]. Available: https://oecd.ai/en/incidents?search_terms=%5B%5D&and_condition=false&from_date=2023-10-01&to_date=2023-12-28&properties_config=%7B%22principles%22:%5B%5D,%22industries%22:%5B%5D,%22harm_types%22:%5B%5D,%22harm_levels%22:%5B%5D,%22harmed_entities%22:%5B%5D%7D&only_threats=false&order_by=date&num_results=20

[3] ‘AI cannot be patent “inventor”, UK Supreme Court rules in landmark case | Law-Order’, OECD, Dec. 2023. [Online]. Available: https://oecd.ai/en/incidents/58264

[4] Wes Davis, ‘Meta disbanded its Responsible AI team’, The Verge, Nov. 2023. [Online]. Available: https://www.msn.com/en-us/money/other/meta-disbanded-its-responsible-ai-team/ar-AA1k9MVq

[5] Cristina Criddle and Madhumita Murgia, ‘Big tech companies cut AI ethics staff, raising safety concerns’, Financial Times, Mar. 2023. [Online]. Available: https://www.ft.com/content/26372287-6fb3-457b-9e9c-f722027f36b3

[6] Rachel Metz, ‘Google is trying to end the controversy over its Ethical AI team. It’s not going well’, CNN, Feb. 2021. [Online]. Available: https://edition.cnn.com/2021/02/19/tech/google-ai-ethics-investigation/index.html

Other Posts

Ensuring AI Readiness: Complying with Article 4 of the EU AI Act

Ensuring AI Readiness: Complying with Article 4 of the EU AI Act

Exploring the Latest in AI Governance and Regulation The EU AI Act is set to revolutionize how organizations engage with artificial intelligence (AI). Central to this legislation is Article 4, which imposes a legal obligation on organizations to ensure a sufficient...

Exploring the Latest in AI Governance and Regulation

Exploring the Latest in AI Governance and Regulation

Exploring the Latest in AI Governance and Regulation Artificial intelligence (AI) is evolving at an unprecedented pace, reshaping various aspects of our lives and society. With its rapid growth, there is a crucial need for robust regulations and policies to harness...

Building a Values-Based Organization: Our Journey Together 🌟

Building a Values-Based Organization: Our Journey Together 🌟

Launching and steering Verdas AI has been a shared experience centered around our mutual values. Here’s a glimpse of how we’ve worked together to incorporate values into our operations, facilitate internal workshops, and align our values with the company’s mission.

Staying Connected Without An Office: The Power of Offsites!

Staying Connected Without An Office: The Power of Offsites!

🌟 Regular Offsites! 🌟

1. Strategic Discussions:
Setting aside dedicated time to discuss strategies, forecasts, and future plans can reignite team alignment and vision. It also provides an opportunity to raise concerns, ask questions, and delve deep into business objectives.