Artificial intelligence (AI) holds the key to greatly improving national security according to some specialists, but it could also potentially cause a great deal of additional problems. Some have suggested that these could include the possibility of serious attacks against major pieces of network infrastructure. Pundits have long predicted hyper-intelligent AI agents that use sophisticated algorithms to calculate the chances of penetrating a particular security paradigm.
Ironically, these might not be the most widespread attacks in the coming years. As government agencies deploy chatbots, there’s a risk that bad actors could confuse them into giving up sensitive information. That’s especially true of any agency that leverages huge amounts of AI-based solutions to handle jobs that were only recently managed by human technicians. A recent document from the executive branch encouraged federal agencies to explore ways they could use AI in their existing workflows. Security experts are concerned about how this could broaden the nation’s attack surface.
Measuring the Overall Network Surface
Representatives from several federal organizations have sought to calculate the odds that any particular AI-related action could impact the amount of exposure the country’s infrastructure experiences. Some agencies, like the Congressional Budget Office, are primarily concerned with the cost of any AI-related programs. These systems certainly have a large upfront price tag that might give tax assessors a real case of sticker shock. Dealing with the aftermath of an attack is extremely costly as well, which is another reason why the CBO is taking a closer look at these issues.
Specialists who work with dedicated public-private partnerships are trying to figure out how many operations certain AI programs could potentially call out per second. There’s a real risk that programs capable of performing actions many times faster than human attackers do could compromise algorithms computer scientists currently view as safe. Hashing functions are a good example.
When online resources store login credentials, they usually hash passwords down to a seemingly random string. Some AI-based tools might be able to decrypt these strings and, therefore, guess the right way to log in faster than a dictionary-based attack ever could. That’s why representatives of these partnerships are once again recommending the use of open-source tools.
Top 5 Epic Final Boss Fights in Video Games(Opens in a new browser tab)
The Rise of Free Software
Non-proprietary security and firewall applications constantly receive updates in a way that conventional programs never could. Those who write commercial code have to deal with deadlines and release schedules. Teams who program open-source competitors recruit enthusiasts from all over the world to revise their code and catch bugs.
Government agencies have been slower to adopt these tools than the general market, but that’s changing. As the risk of AI-based attacks becomes increasingly great, these agencies are reviewing their policies and realizing that there’s no reason not to move away from traditional commercial solutions in certain situations. While it’s true that free software won’t ever completely replace the kinds of boxed products public sector buyers prefer, there’s a good chance that more staffers with decision-making power are going to reexamine their position.
Discussion about this post