tech
Limited Intelligence: How Civil Society is Trying to Take Control of AI
The scale of AI systems implementation in business processes is already such that the problem of algorithm "bias" is seriously affecting society: there are known cases where artificial intelligence has rejected women's resumes during hiring, mistakenly marked black defendants as high-risk recidivists, or denied medical care to the elderly.
TL;DR
- States are failing to regulate AI effectively, leading to ethical concerns and societal impacts like algorithmic bias.
- Civil society is developing non-state mechanisms such as ethical codes, collective testing, media pressure, and audits to control AI.
- Professional organizations like ACM and IEEE are creating ethical guidelines, but these are often not binding.
- Proposed solutions include voluntary rating systems, ethical labeling, bias bounty programs, and community-led audits.
- Ethical licenses (RAIL) aim to prevent the misuse of AI models for malicious purposes.
- Investigative journalism and grassroots activism, including employee protests and user boycotts, play a role in raising awareness and pressuring companies.
- Despite these efforts, the article concludes that fundamental control of AI requires state or international regulation.