How open data can leave gaps for disinformation

Last week’s attack on Iran’s nuclear enrichment facility was reported as a blow to the country’s nuclear programme. It's a significant setback for the country's efforts.

The response came soon enough. Officials announced Iran would boost its nuclear activity and will up the uranium enriching level to 60 per cent purity.

Iran blames Israel for the attack and used the delicate term “nuclear terrorism”. A blackout and sabotage can be catastrophic to the processes of a nuclear enrichment plant, commentators noted.

The New York Times cited anonymous American and Israeli officials claiming Israel was behind the attack. The explosion would have destroyed the independent internal power system, US intelligence officials briefed on the damage said. Mohammed Ayoob, professor emeritus of international relations at Michigan State University analysed for the Australian Strategic Policy Institute (ASPI) the circumstances at Natanz and said several factors point to Israel’s responsibility.

Open data to aid disinformation artists

In this hazy mess of information sharing, disinformation can thrive and quickly spread. One self-proclaimed analyst from Israel cast doubt on the very existence of the attack itself. For the account, no proof was proof enough. On the basis of poor-resolution open satellite data, the handle concluded there were no changes visible.

I went through the satellite images myself and analysed every pixel. I used commercial Planet Lab images, images by Maxar from 5 April (before the attack) and high-resolution synthetic-aperture radar (SAR) images from 13 April, provided by Capella Space. My verdict is that there were no obvious larger-scale changes visible on the surface. But it's no definite judgement on what happened. The resolution is poor, even on these arguable high-resolution images. Important details can still be missed that play a big role. In short, it would be dangerous to make inferential claims on the basis of satellite images alone. 

However, a number of accounts on social networking sites Twitter and Telegram did exactly that and took the details at face value, immediately questioning the whole validity of statements made by Iranian officials. 

“Iran can 'inflate' the extent of the damage in order to breakthrough,” one Israeli Twitter account claimed. Needless to say, the underground facilities could still be damaged without showing any sign of proof on the surface. Last summer, one building caught fire and debris was found distributed far around the building site, a detail that points to an explosion (but then again, we need to be careful not to make claims on the basis of satellite images alone).   

Taking intelligence and making inferential claims is a problem for intentional misleading claims but also for accidental misinformation. Disinformation artists know that half-truths and details that are hard to validate create fertile ground for planting a seed of doubt and fake news. 

Using open-source data may aid the wrong people to build a basis to bolster false narratives. 

There are a number of reasons that might explain why we don't see any signs of change on the surface at the Natanz site. One is that Natanz’s most sensitive work is underground. 

The emergence of disinformation and social media handles casting doubt on media claims is neither new, nor is it surprising. For Natanz, deliberate disinformation is possible, especially in the immediate aftermath of such an incident when information clarity is hazy, writes Gordon Corera, a BBC security correspondent. Disinformation also played a role in last year’s explosion.

Why is the attack in Natanz such a big deal and why is disinformation in this case potentially so harmful to diplomacy? The US and Iran are at a pivotal moment. The two nations were just getting ready to find a path back to a nuclear deal. Although the White House asserted this week that the US had no involvement in the attack, misleading information might help to diminish an important window of opportunity for the US to revive an agreement with Iran - an agreement that Benjamin Netanyahu opposed.

In March, before the attack, the Iranian Foreign Minister Mohammad Javad Zarif urged the US to quickly rejoin the international agreement limiting Iran’s nuclear ambitions, warning that the Islamic Republic’s upcoming elections could stymie progress in any talks.

Where are the limits of open-source intelligence?

The answer is often 'interpretation'. There are recent stories where open data played the leading role in making bold claims. Critics may see it as 'stretching' conclusions. 

In late March, NBC News ran a piece alleging that North Korea had restarted activity at one of its nuclear facilities. On the basis of satellite images, it suggested that Kim Jong Un's regime has started reprocessing plutonium for nuclear weapons. 

The findings are based on smoke plumes emerging from a building's chimney at the Yongbyon Radiochemistry Laboratory site that are visible on Maxar satellite images. Critically, the NBC News reporting left little room for error and alternatives. 

The source of the analysis, Beyond Parallel, a think tank, is much more careful in the claims it makes: "This [the steam (or smoke) arising from any of the stacks], while not an indicator of a reprocessing campaign itself, indicates that the building is occupied and being heated." In short, the facility can be heated for a number of reasons and journalists should make readers aware of that. 

In conclusion, we should ask: should we not benefit from a wealth of new open data and intelligence for interpreting the world? We certainly should, in my view. Should we solely rely on it? Certainly not. Like with many things, there are benefits and risks. Disinformation artists might spin their own narratives and cherry-pick results from open data and satellite images if they find them conveniently matching their narratives. It warrants to be watchful and alert and to double-check with additional sources.