Beyond ethics policies, businesses face an AI accountability test

An AI system can screen thousands of job applications, flag suspicious transactions or recommend which customers should get a loan. But when someone challenges its recommendation, the business using it needs to explain the basis for the result, identify who is responsible and decide whether to correct it.

For Indian organisations expanding their use of AI, these questions are central to how they deploy the technology. The stakes are higher when AI influences decisions about jobs, access to finance or the use of personal data.

The Responsible AI & Ethics category at the ET AI Awards 2027 focuses on organisations putting fairness, transparency, privacy and accountability into practice.

Turning principles into practice

India’s AI Readiness Assessment Report, launched by UNESCO and the Ministry of Electronics and Information Technology (MeitY) at the India AI Impact Summit 2026, highlights this need. Drawing on research and consultations with more than 600 stakeholders, the assessment called for stronger governance, inclusive workforce planning and practical guidance for organisations developing and using AI.

Also Read: When customer service goes beyond the chatbot

For a business, this begins with checking how a system is built and where it will be used. A recruitment team using AI to screen applications, for example, should assess whether the data reflects past hiring biases and whether the system treats different groups fairly. It should also decide when a human reviewer needs to intervene.

Similarly, a lender using AI to assess applications needs to understand the factors behind its recommendations and establish a way to review disputed outcomes. Publishing an AI ethics policy is a starting point, but the policy needs to shape approvals, testing and employee decisions.

Responsibility continues after launch

Checks before deployment cannot account for every situation a system may encounter. New data or changing conditions can affect its performance, making ongoing monitoring and clear procedures for investigating errors essential.

The IndiaAI Mission’s Safe & Trusted AI pillar supports efforts in this area. In an update published on July 27, 2026, MeitY said 13 Responsible AI projects had been approved, covering areas such as reducing bias, protecting privacy, explaining AI outputs, detecting deepfakes and assessing risks.

Also Read: AI Woman Leader of the Year: Making women's leadership visible

Within a business, accountability means assigning a person or team to oversee each system, recording its limitations and setting rules for human review. Employees also need a clear way to flag unreliable outputs and report problems to someone who can act on them.

Where an AI recommendation affects an individual, the organisation should consider how that person can understand the decision and seek a review.

Showing what works

Responsible AI requires technology, legal, risk and business teams to work together. Documented assessments, data safeguards and records of how incidents were handled can help show whether a company’s policies work in practice.

Organisations submitting nominations for the Responsible AI & Ethics category can demonstrate how they assess risks, monitor systems and respond when things go wrong. They can also show who takes responsibility and how their safeguards have improved the way AI is used.

Nominate now for the ET AI Awards 2027.