ai-for-less-suffering.com

← all claims

descriptive claim

Per panelist Stuart Russell, none of the safety activity at the six evaluated companies provides any quantitative guarantee of safety, and the giant-black-box training approach may be structurally incapable of supporting such guarantees.

desc_fli_index_2024_no_quantitative_guarantees

confidence
0.80

Evidence (1)

supports (1)

  • FLI AI Safety Index 2024 expert_estimate
    weight
    0.85

    locator: Panellist Comments (Stuart Russell)

    “none of the current activity provides any kind of quantitative guarantee of safety; nor does it seem possible to provide such guarantees given the current approach to AI via giant black boxes”

Camps holding this claim (3)