Marc Pritchard made an important call to arms recently with his challenge to digital advertising to clean up its act. His charge, that we have embraced all the benefits of digital advertising without really addressing the downsides, requires every serious player in the industry to respond.

The issue is not just that digital advertising needs to heed the same standards as its more established siblings. The deeper issue is that the digital advertising model is increasingly audience-driven — we buy custom audiences either directly or through automated exchanges, without making a distinction about where those audiences are found.

As we shift to a media model that is purely audience-driven, and where editorial context is not considered, we risk placing brands in environments where they can come to serious harm.

The most glaring example of this harm was the discovery by the Jaguar marketing clients in the UK that their ads were running next to ISIS recruitment videos on YouTube. This has led to Jaguar suspending its entire digital advertising budget for a day. A recent Times of London investigation found that many clients in the UK have their brands in unsafe places, from a brand association perspective.

One of our UK clients had digital ads running on a neo-Nazi site. This was not intentional, of course, and was quickly stopped. These ads were bought programmatically and simply followed our audience to the darkest corner of the Internet.

So we find ourselves at a crossroads.

The benefits of taking an audience-driven approach in advertising are profound: more relevance, less waste, and better transactional results. But the risks to brand health and safety are becoming too big to ignore. Of course, no one should believe Jaguar is looking to attract ISIS followers, but one screenshot of that ad next to a beheading and the grisly association is permanent.

And the scale and exploding growth of the digital advertising ecosystem means we need to get serious — and quickly.

What are we to do? At UM, we are implementing, immediately, SAFE (Sites Automatically Filtered & Excluded), a brand safety blacklist of wsites that will be excluded from all digital media buys, programmatic or otherwise, for all our clients. This list includes all the obvious candidates based on topic – sites promoting racism, anti-Semitism, neo-Nazism, homophobia, misogyny, violence, terrorism, pornography – and others based on intent – fake news sites, clickbait farms, bots.

All our digital media buying platforms and partners will have to verify these Web sites are not part of any client media buy. Further, UM will not rule out blacklisting ad tech providers whose code appears on sites on the SAFE blacklist, too.

The SAFE blacklist will be automatic for all our clients, which means they would have to put in writing to UM that they don’t want to be part of this program — and then opt out. We would try to talk them out of opting out, but ultimately we are agents to our clients.

Why haven’t we done this before? We have blacklists for many of our clients currently, but they are ad hoc and have been built up over time reactively in response to bad experiences. They aren’t systematic nor comprehensive. The SAFE blacklist will combine all these ad hoc blacklists into one agency-wide

The blacklist that will be updated formally by our senior digital leadership and shared with our clients every month.

Does this approach require us to make values-based judgment calls? Yes. We do it every day in media. It is what we do when we decide to run J&J ads on morning TV and Coca-Cola ads in the Super Bowl and Chrysler ads in late-night comedy. Our judgment is why clients come to us in the first place and judgment about the right context or moment for their brand is the primary advice that clients are looking for from us.

How will we verify that our SAFE blacklist has been avoided? We will work with all major third-party verification partners we use: IAS, MOAT, WhiteOps and DoubleVerify. All will participate in the maintenance of the list and in verification against it.

How will the SAFE blacklist deal with recent Google/YouTube incidents? In the case of YouTube, our approach is to direct advertisers to more brand safe inventory (YouTube Masthead, Google Preferred) and to work with clients to pressure Google to create its own internal SAFE blacklist and to filter all content on their platform (not just the premium content).

We will also continue to pressure Google to open YouTube fully to independent third-party verification.

Can anyone see our blacklist? Yes. We will be sharing SAFE with anyone in the industry who asks to see it. We have nothing to hide. As a media industry, we need to have nothing to hide.

What about clients who want a Whitelist? We have clients who have gone further and implemented a Whitelist of sites where their brands can only ever be placed. We will honor these and ensure they are similarly held up to verification.

At UM, we have just launched our ninth report-out from our global Wave study, which looks at global social and digital behavior around the world. In it, we discovered that 48% of people on the Internet, accessing mostly through smartphones, feel stressed and overwhelmed by the sheer volume of content.

It is fair to say  this stress is shared by marketers, doubly, as people and as brand stewards.

We have a unique opportunity now, as digital advertising outpaces TV spend, to protect the amazing advantages of the digital brand experience, while eliminating the risks we are only now beginning to fully understand. As an agency, it means taking a hard look at everyone in the digital space, especially ourselves, and protecting the brands of our clients better, which we have an obligation to do.

Recent times have shown the online world is darker than we thought. SAFE is our proactive way of keeping ourselves, our clients and their brands out of harm’s way. And if it means ISIS recruitment videos, neo-Nazi sites and fake news sites get less digital advertising dollars, then so be it.