TV news does need better rules on AI and shouting matches. A vague, unnamed warning from the ministry is the wrong way to write them

Anyone who has watched an Indian prime-time debate knows the problem the Ministry of Information and Broadcasting is pointing at. Panellists shout over each other, anchors egg them on, and remarks that would be unprintable in a newspaper go out live to millions. Now channels have started adding virtual presenters and AI-made graphics, sometimes with little sign that a machine was involved.

On Wednesday the ministry told private satellite channels to follow the Programme Code strictly. It said an inter-departmental committee had found telecasts not in keeping with the code’s “spirit”, and noticed near-identical reports and visuals running across several channels. It told broadcasters that the use of AI “may be limited to assistive capacity”, asked them to screen and brief debate panellists, and to mute, warn or cut the feed when speakers turn abusive. Breaches, it said, could invite regulatory action.

The fair point

Much of this is sensible. A channel that lets a guest spew communal abuse is not exercising free speech; it is choosing ratings over its duty to viewers. The worry about AI is real too. A synthetic anchor reading a script, or a generated image presented as footage, can mislead viewers in ways that are hard to spot and harder to correct. Newsrooms should keep human editors responsible for what goes out.

Where the advisory goes wrong

The trouble lies in how the ministry has gone about it.

The advisory names no channel and no programme. A broadcaster cannot learn what it did wrong, and the public cannot judge whether the committee was right. Warnings without specifics work less as regulation than as a general chill, and they leave room for the suspicion that the target is coverage the government dislikes rather than conduct that harms viewers.

The language is loose where it should be tight. “Assistive capacity” is not a rule a newsroom can follow; it does not say whether an AI-generated map, a translated caption or a synthetic voice reading the weather is allowed. And asking channels to screen out “anti-national” content in guidance on debates hands an elastic political label to the very ministry that licenses them.

A better way

First, the ministry should publish the advisory in full and, where it has found breaches, name them and give channels a hearing through the grievance system that the cable television rules already provide.

Second, on AI, the test should be honesty, not a blanket limit. Any AI-generated presenter, voice or image should carry a clear on-screen label, and nothing synthetic should ever be passed off as real footage. Broadcasters’ own self-regulatory bodies should draw up the detailed code, with the ministry stepping in only when they fail.

Third, the standards on abuse and incitement should be the ones already in law and in the Programme Code, applied evenly to every channel, whatever its politics.

Better television news is a goal the government and viewers share. It will come from rules everyone can read and apply, not from warnings passed quietly to broadcasters.

The reporting behind this view

• 0channels or programmes named in the ministry's advisory of 21 October
• 'Assistive'the only limit the advisory sets on AI in news and programmes
• 1995year of the Cable Television Networks Act under which the Programme Code sits