Aug 31, 2026 · 1h 1m
Listen from 10:54
Listen at 10:54
Deepfakes turn the internet’s trust crisis into enterprise risk
AI Has Made Everything Worse | Hany Farid on the Internet's Truth Problem
Synthetic media is moving beyond misleading posts into impersonation, fraudulent hiring, and attacks that can compromise companies and institutions.
- 1Social platforms rewarded engagement over truth before generative AI made manipulation cheaper and harder to recognize.
- 2Deepfakes now threaten enterprises through executive impersonation, voice cloning, fraudulent hiring, and North Korean IT-worker infiltration.
- 3Forensic detection, clear organizational ownership, platform accountability, and regulation offer practical defenses despite worsening near-term risks.
Don't miss
Farid describes North Korean IT workers infiltrating major organizations, turning deepfake and identity threats into immediate security problems.
The brief
Hany Farid argues that generative AI did not create the internet’s truth problem; social platforms had already optimized for attention, outrage, and incompatible realities.
The threat is now operational: fake executive calls, cloned voices, fraudulent identities, and North Korean IT workers are penetrating companies and defense contractors.
Farid says individuals cannot become reliable forensic analysts overnight, so platforms, regulators, and organizations must carry more responsibility for reducing manipulation.
Detection requires layered forensic analysis rather than a magic enhancement tool, and companies need training, phishing awareness, and clear ownership of deepfake risks.
Farid expects conditions to worsen before defenses mature, but sees institutional learning, better detection, and accountable governance as reasons for cautious optimism.