AI tools are entering newsrooms with claims they’ll revolutionize journalism. The reality is more mundane. AI helps with specific tasks but doesn’t replace core journalism skills. Here’s what’s actually happening beyond vendor marketing.

What AI Actually Does

AI tools in journalism fall into a few practical categories that are useful without being transformative.

Transcription: Speech-to-text has become accurate enough to transcribe interviews automatically. Journalists still need to review for errors and context, but it saves hours of typing. This is genuinely useful.

Research assistance: AI can scan large document sets looking for patterns or specific information. When you have thousands of pages of financial records or government documents, AI tools help identify relevant sections faster than manual review.

Translation: Neural translation has improved significantly. Journalists covering international stories can get workable translations of foreign language sources quickly. The translations need review by humans who know the language, but they provide starting points.

Data analysis: For stories involving datasets, AI can identify patterns, anomalies, or relationships. A journalist still needs to understand what the patterns mean and whether they’re newsworthy, but the AI handles the computational heavy lifting.

These applications are real and save time. But they’re assistive tools, not replacements for journalism.

What AI Can’t Do

The core journalism work—finding stories, conducting interviews, evaluating credibility, providing context, writing with clear voice—remains human work.

Judging newsworthiness: AI can’t determine what’s important. It can identify that something is unusual or different from historical patterns, but deciding whether that’s newsworthy requires human judgment about what matters to audiences.

Understanding context: An AI might notice that a politician’s statement contradicts a previous statement, but understanding whether that contradiction is meaningful requires knowing the political context, the person’s track record, and how audiences interpret such contradictions.

Source evaluation: Journalism relies on assessing source credibility. Is this source reliable? Do they have agenda? What’s their track record? AI tools can check factual claims against databases, but evaluating human sources requires human judgment.

Ethical decisions: Every journalism decision involves ethics. What to report, how to report it, whose voices to include, what might cause harm. These decisions require values and judgment AI doesn’t possess.

The Automated Content Problem

Some outlets use AI to generate routine content: sports recaps, earnings reports, weather updates. This works for formulaic content where the structure is predictable and facts come from structured data.

These automated articles are fine for basic information delivery. They’re not journalism in any meaningful sense. They’re automated templating.

The risk is that organizations see this working for routine content and assume it will work for actual reporting. It won’t. Original reporting requires investigation, interviews, judgment, and narrative craft that AI can’t replicate.

Fact-Checking Assistance

AI tools help fact-checkers by flagging claims that contradict known facts or lack supporting evidence. This speeds up identification of potential false claims.

But verification still requires human work. AI might flag that a claim about unemployment rates doesn’t match official statistics. A fact-checker must then investigate whether the discrepancy reflects data interpretation, timing differences, or actual falsehood. Context matters, and AI doesn’t provide it.

SEO and Audience Analytics

Media organizations use AI extensively for SEO and understanding audience behavior. These tools analyze what headlines perform well, which topics drive traffic, and how to optimize content for search engines.

This optimization is useful but creates incentives toward clickbait and sensationalism. AI optimizes for engagement metrics, not journalism quality. These goals sometimes align but often don’t.

The risk is that AI-driven optimization pushes journalism toward whatever generates clicks rather than what serves public interest. This tension existed before AI but becomes more acute when algorithmic recommendations drive editorial decisions.

The Economics

Media companies face financial pressure. AI promises cost reduction. This leads to two concerning trends:

Headcount reduction: Some organizations reduce journalism staff while implementing AI tools, claiming AI compensates for lost capacity. It doesn’t. You can’t replace experienced reporters with software.

Content volume emphasis: AI enables producing more content faster. But journalism isn’t manufacturing. More articles don’t equal better coverage. In-depth reporting requires time and expertise that can’t be rushed or automated.

Detection and Trust

As AI-generated content becomes common, audiences struggle to distinguish human journalism from automated content. This erodes trust in all media.

Some publications clearly label AI-assisted content. Others don’t, letting readers assume everything is human-written. This lack of transparency damages credibility.

The solution requires industry standards for disclosure. If AI was used in reporting or writing, say so. Let readers make informed judgments about what they’re reading.

The Investigative Journalism Gap

Complex investigative work—examining corporate malfeasance, government corruption, systemic issues—requires extended research, cultivated sources, and deep expertise. AI tools might assist with document analysis, but the core investigation remains human work.

There’s concern that as media organizations invest in AI for routine content, resources for investigative journalism shrink. The economics push toward cheap, voluminous content over expensive, time-intensive investigation.

This matters because investigative journalism provides accountability that nothing else replaces. If media economics make this kind of work unsustainable, we lose something critical regardless of how efficient AI makes routine content production.

Skills Journalists Actually Need

AI changes what skills matter for journalists:

Data literacy: Understanding how to work with datasets, interpret statistical analysis, and use AI tools effectively becomes more important.

Critical evaluation: As information becomes abundant, the ability to evaluate credibility and importance becomes more valuable than information gathering itself.

Technical understanding: Journalists covering AI, algorithms, and technology need deeper technical knowledge to ask informed questions and identify misleading claims.

Narrative craft: Writing skills remain essential. AI can produce text, but compelling narrative that engages readers and explains complexity clearly requires human craft.

What Newsrooms Should Do

Implement AI tools where they genuinely help: transcription, translation, data analysis. These applications free journalist time for work that requires human judgment.

Don’t use AI to replace journalists. Use it to make journalists more effective. The value proposition is efficiency, not headcount reduction.

Establish clear policies on AI use and disclosure. Audiences deserve to know when and how AI was involved in content production.

Invest in training so journalists understand AI capabilities and limitations. They need to use these tools effectively without overestimating what they can do.

The Bigger Picture

Technology has always changed journalism. The printing press, telegraph, radio, television, internet—each shifted how news is gathered and distributed. AI is another shift, not the end of journalism.

The core journalism functions—investigating, verifying, providing context, holding power accountable—remain human work. Technology changes how journalists do this work but doesn’t eliminate the need for it.

The risk isn’t that AI replaces journalism. It’s that economic pressure uses AI as excuse to underfund journalism while claiming efficiency gains compensate for reduced capacity.

Looking Forward

AI in journalism will expand. Tools will improve. Some applications will prove valuable. Others will be abandoned as their limitations become clear.

The journalism that survives will be work that requires human judgment, expertise, and craft. Commodity content will increasingly be automated. Original reporting, investigation, and analysis will remain human.

Media organizations that recognize this distinction and invest accordingly will maintain relevance. Those that chase automation and cost-cutting will produce content but not journalism.

The difference matters.