Interview with IBM's Executive VP John Kelly III on the company's plans to build ethical AI systems during and after the COVID-19 pandemic
Linda Kinstler / Protocol : Tweets: @protocol and @mcwm . Thanks: @mcwm Tweets: @protocol : IBM has stepped in to fight the coronavirus pandemic. Among the things it's doing is working on ways to track employees to stem the spread of the virus. But don't worry: It's opt-in, and the data is dumped after a couple weeks. https://www.protocol.com/... Mike Murphy / @mcwm : This is really interesting: IBM is has an (opt-in) app for its employees in India to track their movements and potentially learn how COVID-19 is being spread https://www.protocol.com/... Thanks: @mcwm
Context & Ripple Effects
IBM is using the pandemic as a proving ground for the ethical-AI brand John Kelly III has been articulating since the company's 2018 study found ~60% of polled execs worried about explaining AI use to regulators. The concrete test is an opt-in app tracking IBM employees in India to trace COVID-19 spread, with data deleted after a couple of weeks — a deliberate contrast with the surveillance-style contact tracing other employers were adopting.
The interview matters because IBM is simultaneously selling pandemic AI into enterprises: within months, budget-cut call centers were turning to Watson chatbots to filter calls, and IBM was building return-to-office decision tools alongside ADP. Kelly's ethics framing is the pitch that keeps those deployments politically sellable.
First-order effects
- IBM employees in India get an opt-in tracking app whose stated safeguard — deletion after roughly two weeks — becomes the template IBM must defend publicly for every pandemic deployment that follows.
- Kelly's interview forces IBM to hold the ethics line while its own commercial units push Watson into cost-driven replacements for human call-center operators.
Second-order effects
- The tracking-app pattern extends from virus tracing to workplace management: IBM and ADP are already developing tools to automate decisions about bringing employees back to the office, moving employer surveillance from health emergencies into ordinary HR operations.
- Rivals selling workplace analytics and chatbots now face an ethics benchmark IBM is setting — opt-in consent and data expiry become table stakes in enterprise pitches rather than differentiators.
Third-order effects
- If the pattern holds, pandemic-era employee monitoring normalizes into standing operational AI governance, with explainability — the concern IBM documented back in 2018 — as the regulatory hinge; the arc runs from Kelly's 2020 framing to CEO Arvind Krishna later conceding that Watson's health care push was 'inappropriate', showing how hard the ethics-to-deployment gap is to close.
The trend: Enterprise AI vendors are converting pandemic-era crisis deployments into permanent workplace governance products, with consent and explainability as the compliance currency.